Big Data 2.0: cataclysm or catalyst?

1. Four undeniable trends shape the way we think about data –
big or small. While managing big data is ripe with challenges,
it continues to represent more than $15 trillion in untapped
value. New analytics and next-generation platforms hold the key
to unlocking that value. Understanding these four trends brings
light to the current shift taking place from Big Data 1.0 to 2.0,
a cataclysmal shift for those who fail to see the shift and take
action, a catalyst for those who embrace the new opportunity.
Whitepaper
BIG DATA 2.0
Cataclysm or Catalyst?
A Leader’s Guide to Navigating the Shift in Big Data
www.actian.com

2. BIG DATA 2.0 – Cataclysm or Catalyst?
www.actian.comWhitepaper
The world has changed.
We now live in the digital age. Everything is made up of bytes and pixels. Billions of devic-
es, sensors, and apps create a constant flow of data. All things digital are now connected
by the internet of things. People now use devices with hundreds of apps utilizing all of
the sensor technology embedded in the device and creating new data and new patterns
just waiting to be discovered.
The world will keep changing.
The only thing we know to be constant is change. Two elements drive unprecedented
change. First, the entire world is in flux. We are in unchartered territory and no one really
knows what might happen next. Second, technology innovation is exploding on four vectors
all at the same time. The internet connects everything, everywhere. Mobile device adop-
tion outpaces traditional desktop and laptop computing. The cloud is taking over the cor-
porate data center. And big data continues to explode. Change will no longer be measured
by seasons or eras, it will happen constantly at a much faster pace than ever before.
Time is the new gold standard.
As the speed of change continues to accelerate, time becomes the most valuable commod-
ity. You can’t make more of it, you can only move faster or use it more wisely. Everyone
expects instant access and immediate response. Speed becomes an unfair advantage to
those who possess it. Anyone who wants to survive in the new age must combine speed
with intelligence in order to come out a winner.
Data is the new currency.
You’ve heard it said that data is the new oil, but we say that data is the new currency.
He who has the data wins in the end. With the dropping cost of storage and new technol-
ogy designed to capture, store, and analyze data at ever decreasing costs, companies are
holding on to more data than ever before. Hidden in the data are the secrets to attaining
more customers and keeping them, shaping their behavior, and minimizing enterprise risk
at levels never before possible. The combination of massive amounts of data and new ana-
lytic algorithms creates a new global currency, more stable than the dollar or the pound or
the yen.

3. BIG DATA 2.0 – Cataclysm or Catalyst?
www.actian.comWhitepaper
Big Data 1.0 – A Laboratory Science Experiment
It’s in the midst of these four trends that we have exploded into the age of big data.
After years of big data hype everyone wants to know the state of the union. Big Data 1.0
has been all about introducing new technologies to take advantage of all the new data
being created. One of the frontrunners that has emerged with great promise is Apache™
Hadoop.
From a 50,000 foot view, Hadoop has gone from being an open source phenomenon like
no other to becoming a common household name among technical and business people
alike. Big data has evolved into something beyond what we would call a market; it’s a
movement, a global movement.
Enormous, affordable scale compared to old storage paradigms
The traditional storage market has been under attack from above and below. From above,
cloud vendors like Amazon, Google, and Microsoft use cloud storage volumes to drive down
the overall cost of simple storage. From below, HDFS (Hadoop Distributed File System)
now offers a simpler way of storing data than the higher cost options provided by NAS,
SAN, and databases.
Capture and store all data without first determining its value
With the cost of storage at bargain prices, everyone from the phone company to the gov-
ernment decided to keep data that had always thrown away. Energy companies keep their
historian data. Log files mount up and never before rates. And new data from sensors,
social media, and the internet of things has found its way onto Hadoop storage systems
everywhere.
New types of data present new opportunities for analytics
All of the new data being amassed in Hadoop represent new opportunities for analytics.
Mobile apps produce huge amounts of behavioral data that can be used to drive hy-
per-segmentation. Social data provides never before captured context to what is happening
around markets, brands, and segments. And sensors create micro-snapshots that can be
pieced together into valuable predictive models.
Data discovery and data provisioning now common
in most organizations
New storehouses of data pouring into file system storage have spawned new practices
around data discovery and data provisioning. The first order of business around big data
has been to discover what is actually in the data. The next order has been to provision that
data to the different people and applications that need the data.
Analytic applications emerge as the number one use for big data
For everyone involved in the big data movement there seems to be consensus around what
to do with the oceans of data that have been collected and categorized. Everyone under-
stands that analytics unlocks the value within the data.
The growth and adoption of Apache™
Hadoop continues to astound the market-
place. From its first release in December of
2007 until now there widespread distribu-
tion of the new data platform has exploded
thanks to the open source community and
companies like Hortonworks, Cloudera, and
MapR. This first phase of big data brings
us 5 significant advancements that change
the way organizations all over the world
approach data and analytics:
1. Enormous, affordable scale compared
to old storage paradigms
2. Capture and store all data without
first determining its value
3. New types of data present new
opportunities for analytics
4. Data discovery and data provisioning
now common in most organizations
5. Analytic applications emerge as the
number one user of big data
5 Advancements of Big Data 1.0

4. BIG DATA 2.0 – Cataclysm or Catalyst?
www.actian.comWhitepaper
Extremely complex to integrate, deploy, operate, and manage
The Hadoop open source organization has spawned a myriad of component projects being
run by diverse groups all over the world. It is difficult, at any given time, to determine
which of the projects will graduate into the code base and which will fail. The complexity
of the pieces parts combined with the diversity and explosion of data make integrating,
deploying, operating, and managing massive Hadoop environments difficult. There is no
such thing as an out-of-the-box big data solution.
Requires specialty programmer skillsets to deploy the system
and process data
Many companies who initially stood up their Hadoop clusters thinking they were going to
run predictive analytics on their new platform have abandoned that pursuit. When they
saw the amount of MapReduce, Java, or Python programming that it was going to take to
deploy the system, they scaled back their effort, focused Hadoop on data provisioning,
and brought in analytic platforms to do the heavy lifting analytics. There are not enough
resources available in the world to fulfill everyone’s big data dreams and that is not going
to change any time soon.
Lacks enterprise-class capabilities like security and high availability
Even though big data platforms have sprung up in most companies, they have stayed out
of production because of the difficulty of protecting the data and keeping the systems up
and running. Even big companies like Google and Facebook, that run their operations on
Hadoop, won’t let anyone or anything near the clusters for fear they will bring the entire
system down. While both of these issues are being addressed by the open source commu-
nity and other vendors, they remain a barrier to production.
Shortage of data science skills
The top performer in the world of big data is the data scientist – the person who mines
all the data to find the valuable nuggets to solve specific business problems. However, it
is such a unique skillset that very few people actually possess what it takes to succeed
in this role. Success takes a combination of mathematical expertise, business acumen,
understanding of the new data types, out of the box thinking, and storytelling. Multiple
studies, including the recent study done by McKinsey, have concluded that there will
continue to be a shortage of data scientists.
Performance lags behind the need for immediate response times
Because Hadoop is a file system built on batch processing, wait times continue to be too
long for the operational needs of most businesses. To compound matters, MapReduce was
designed to search and find data, not to run the sophisticated analytics that companies
expect from their big data implementation.
Even with all of strength of a global
movement, most big data projects are
stuck in the laboratory. It is still a science
experiment in need of production. Millions
have been invested and the executives
behind the funding are still waiting to see
a return on their investment. So, what is
holding back the advancement of big data?
Big Data 1.0 unearthed 5 big challenges
that need to be overcome in the next wave
of the movement:
1. Extremely complex to integrate, deploy,
operate, and manage
2. Requires specialty programmer skillsets
to deploy the system
and process data
3. Lacks enterprise-class capabilities like
security and high availability
4. Shortage of data science skills
5. Performance lags behind the need for
immediate response times
5 Challenges of Big Data 1.0

5. BIG DATA 2.0 – Cataclysm or Catalyst?
www.actian.comWhitepaper
Cooperative processing will deliver faster time to value
and better price performance.
The concept of cooperative analytical processing began to emerge in Big Data 1.0. In Big
Data 2.0 it will become the norm. Everyone agrees that there is no one single platform
that can handle all the workloads necessary to run today’s complex organizations. Running
workloads on platforms that weren’t designed with that genre in mind will always cost
more money. When companies implement Hadoop for data management and provisioning,
traditional databases for operational workloads, high performance analytical databases for
predictive and prescriptive workloads, they will naturally drive down the cost of computing
and accelerate time to value. When these new systems support cooperative processing
across the different platforms they will extract even more value from the data.
Analytic building blocks will provide accessibility for
non-skilled and less-skilled workers.
With the shortage of data scientists written in the books, Big Data 2.0 opens up the world
of big data analytics to non-skilled and less skilled workers. Analytic functions delivered
through visual interfaces or embedded in databases become building blocks accessible to
anyone via simple SQL calls or a drag and drop work bench. Users no longer have to know
how to write sophisticated algorithms; they just need to know how they work and how to
use them.
Moving processing to the data will operationalize big data
and push toward real-time.
It is the concept of massively parallel processing that drives the operationalization of big
data. Up until recently, only massively parallel analytic databases were able move the pro-
cessing to where the data lives, right on each node. Platforms like Actian Analytics Plat-
formTM optimize and then compile the query to run on hundreds of compute nodes. With
the introduction of YARN (Yet Another Resource Negotiator) introduced as part of Hadoop
2.0, products like Actian Accelerator for Hadoop now run sophisticated data transforma-
tions and complex analytics directly on the HDFS (Hadoop Distributed File System) node.
This major breakthrough puts hundreds of Hadoop nodes to work and greatly accelerates
the speed of computing.
Combining non-relational and relational data will
enable a richer set of analytics.
The first slide of the Facebook presentation at the 2013 Strata conference in New York
City stated, “Big Data Does Not Equal Hadoop.” The second slide stated, “Big Data
Equals Hadoop plus Relational.” When a top contributor to open source Hadoop put a
stake in the ground around the need for both of these two platforms together, the world
listens. Companies that want to excel in their industries and utilize their data in the most
profound ways possible will implement new technologies that work together across both
relational and non-relational data.
Big Data 2.0 is all about releasing the $15
trillion of value from below the water line
of the big data iceberg. The race is on to
provide affordable access to the 88% of the
data that has never been touched. To move
quickly to the transition to Big Data 2.0,
you can expect six major shifts on the tech-
nology front to give companies the ability
to move quickly, turn on a dime, and stay
ahead of the competition, even in the face
of relentless change:
1. Cooperative processing will deliver
faster time to value and better price
performance.
2. Analytic building blocks will provide
accessibility for non-skilled and less-
skilled workers.
3. Moving processing to the data will
operationalizes big data and push
toward real-time.
4. Combining non-relational and relational
data will enable a richer set of analytics.
5. Services layers will abstract away the
complexity of underlying infrastructure.
6. Unified platforms will provide modular
approaches covering the entire analytic
process
Big Data 2.0 – Transformational Value

6. BIG DATA 2.0 – Cataclysm or Catalyst?
www.actian.comWhitepaper
Better price performance
Expect newer technology to drive down the cost of computing. Older, traditional technology
was not designed to handle diverse data or to crunch through huge data analytics. Because
modern platforms run on commodity hardware and were designed for the data deluge they
will cost you less and perform better. Get out your calculator or spreadsheet. If you don’t
see better price performance, don’t buy it.
Faster time to value
By abstracting away complexity and speeding the processing of data preparation and ana-
lytics, you will accelerate your time to value. You’ll cut data processing times in half and
speed analytics iterations by magnitudes of more than 100. Not only will you get value
out of your data faster, you’ll be able to deploy new data services and analytic applications
faster than ever before.
Continuous transformation
To keep up with the change around us, it’s no longer enough to be able to transform your
company. The speed of Big Data 2.0 will give you the innovation acceleration you need
to transform your company over and over again. You will be able to use analytics different
than your competition, disrupt your market, and lead your industry.
Big Data 2.0 Means Big Value Times Ten
So, what does all this mean for executives
and companies who have invested in big
data technology or who have an investment
in their 2014 plans? The combination of
capabilities being deployed in Big Data 2.0
should yield four points of distinct value for
those who are willing to lead the charge:
1. Better price performance
2. Faster time to value
3. Continuous transformation
4. Unfair advantage
Services layers will abstract away the complexity of underlying
infrastructure.
One of the greatest challenges of Big Data 1.0 was the complexity of data types and plat-
forms required to handle all of the different workloads processed by most organizations.
When you add in the complexity of running analytics using extreme SQL, Java program-
ming, MapReduce, Python and other scripting or NoSQL languages, the complexity can be
overwhelming. In Big Data 2.0 the complexity will be abstracted away using new technolo-
gy like extended optimizers, common metadata, and visual frameworks that make big data
analytics accessible to the masses.
Unified platforms will provide modular approaches covering the
entire analytic process.
New unified platforms designed for Big Data 2.0 and beyond will cover the entire analytic
process from connecting, preparing, and enriching data all the way to applying and com-
bining sophisticated analytical functions. To support the protraction of big data projects
and the growth of data over time, these new platforms will deploy in modules, growing by
function, node, or data storage volume. In addition, expect modern data platforms to de-
ploy on-premise, in the cloud, or in hybrid environments, all on commodity hardware. This
modular approach will give way to adaptive infrastructure, where nodes can be deployed
on one platform and quickly spun down and spun up to support a workload increase on
another platform.

7. BIG DATA 2.0 – Cataclysm or Catalyst?
www.actian.comWhitepaper
Unfair advantage
All of this creates an unfair advantage for your organization. When you combine better
price performance and faster time to value with continuous transformation, no one will be
able to catch you. You will become the next Amazon, the next Google.
Transformational Value
So, what makes next-generation technology transformational? It’s the extreme speed,
extreme scale, and extreme agility that technological advances bring to the market place.
Truly innovative companies are moving beyond the confines of old technology and first
generation big data platforms to embrace Big Data 2.0.
Think of technology like Actian Analytics Platform as the exoskeleton for Hadoop. Hadoop
captures all the data. It becomes the data operating system. It prepares and provisions
data for anyone, anywhere. It enriches data sets with fresh perspectives turning ordinary
analytics into a strategic weapon. Actian surrounds big data open source components with
extreme performance, extreme scale, and extreme agility. It provides access to less skilled
and non-skilled contributors. It opens the door to the $15 trillion of value hidden below
the water line.
In order to succeed in the new world that is continuously emerging, we need to do three
things: accelerate everything, innovate constantly, and transform continuously.
Accelerate everything.
Every step of data’s journey through next-generation platforms must be accelerated. We
must leave no stone unturned when it comes to faster processing, better optimization, and
quicker human interaction with data and machines.
Innovate constantly.
Since the only constant is change, we must constantly innovate. A single innovation will
only set us apart or put us in the lead for a short period of time. We need more innovation
on the tail of every successful innovation we take to market.
Transform continuously.
The old paradigm of enterprise transformation has given way to the new model of continu-
ous transformation, where there is always some part of the organization being transformed.
It many cases new transformation cannibalize old parts of the business, but this is neces-
sary to stay alive and to stay ahead.

8. BIG DATA 2.0 – Cataclysm or Catalyst?
www.actian.comWhitepaper
It’s Business Time!
With next-generation technology emerging to match
the explosion of big data, it’s business time. It’s time
for the business to reap the benefit of their big data
investments. Visual frameworks combined with drag-
and-drop operators are ready to plug into any big
data reservoir and produce immediate value. High-
performance analytics platforms are poised to provide
accelerate the proliferation of analytics across entire
ecosystems. Big Data 2.0 will be the catalyst for the
next wave of Amazon’s, Google’s, and Facebook’s.
It’s business time for Big Data 2.0.
Actian transforms big data into business value for any organization – not
just the privileged few. Our next generation Actian Analytics Platform
software delivers extreme performance, scalability, and agility on off-the-
shelf hardware, overcoming key technical and economic barriers to broad
adoption of big data, delivering Big Data for the Rest of Us. Actian also
makes Hadoop enterprise-grade by providing high-performance ELT, visu-
al design and SQL analytics on Hadoop without the need for MapReduce
skills.